Financial inclusion and the contested infrastructures of cash transfer payments in South Africa

Introduction

In Mahlathi, a rural village in South Africa’s northern Limpopo province, community activist Israel Nkuna has helped thousands apply for the COVID-19 Social Relief of Distress (SRD) Grant, launched by the government in March 2020 following stringent public health lockdowns. The grant was the first to target unemployed, working-age people who did not receive other forms of state income support. For Nkuna, however, ‘the grant hurts people instead of helping them. It gives us hope that we’ll eat again and then takes it away in a bureaucratic mess’ (personal communication, 2022). Of those 13 million who initially applied for the grant, only 8 million were approved (Mafata, 2021). Every day, Nkuna receives hundreds of messages from applicants dealing with delays, rejections and the frustrations of dealing with online and mobile registration systems that seemingly deny them access to their constitutional right to social protection. Digital technologies are becoming increasingly common in the provision of cash transfers, particularly in the Global South (Devereux and Vincent, 2010; Masiero and Prakash, 2019; World Bank, 2022). They are frequently pitched as efficient solutions to the problems of distribution in environments where cash is risky and financial infrastructures are limited. We know very little, however, about how these systems are experienced by recipients, how they mediate the provision of social protection and, more broadly, what they mean for social citizenship.

In South Africa, 31% of the population receives some form of state income support, known as social grants – a number that swelled to nearly 50% during the expansion of social protection spending during the COVID-19 pandemic (SASSA, 2022). In the context of chronic unemployment and limited job growth, the expansion of social grant spending has become a critical income source for millions of households. The distribution of social grants in South Africa has long relied on a mix of public and private channels, including public utilities, private banks and financial technology firms. In the last decade, and mirroring global poverty management trends, mobile and financial technology companies have become increasingly dominant to the registration and distribution of income supports. During the COVID-19 pandemic, for example, the South African government contracted a private technology firm to develop mobile platforms for the payment of some 10 million grants. Around the same time, the World Bank and Gates Foundation rolled out their updated Government to Persons Payment systems approach (G2Px), which further entrenches the role of financial technologies in the targeting and distribution of state income transfer programmes. As Gabor and Brooks (2017) have argued this ‘fintech-philanthropy-development complex’ (FPD), is increasingly central to the management of poverty and the development of policy agendas across much of the Global South.

In this article, we examine how these digital and financial technologies affect recipients’ experience of receiving social protection in South Africa. We are particularly concerned with how these technologies are shaping social citizenship and mediating relations between citizens and the state. At a global scale, we build on work which has noted how digital and biometric technologies are reshaping the delivery and experience of social protection in the South (Carswell and De Neve, 2022; Chaudhuri, 2021; Raphael, 2022). Rather than neutral technologies which promote distributive efficiencies, these interventions provoke anxiety, confusion and unease among grant recipients. The melding of social protection measures with financial technologies has created new informational opacities and new forms of exclusion. Building on existing work on the changing political economy of cash transfer programmes in the Global South (Ansari, 2022; Ballard, 2013; Castel-Branco, 2021; Cookson, 2016; Ferguson and Li, 2018; Green, 2021; Torkelson, 2020), we argue that these distributive technologies are not only reshaping the experience of receiving social protection, but the nature of social citizenship itself. As Mader (2018) has argued, the logics of financial inclusion have opened up space for a range of private actors to shape development policy – and their rationale for doing so is never entirely in the public interest. Rather than opportunities to develop more universal income support programmes, these technologies are aimed at developing targeted and temporary forms of emergency welfare support that can be rolled out and then retracted when the crisis ends. Across much of the Global South, market participation has come to dominate social policy and people’s livelihood strategies (Nguyen et al., this volume). Amid inadequate wages and social protection, financialized debts are increasingly used to address income gaps which increase risk and uncertainty. This has significant implications for how we understand welfare policy, as well as political subjectivities and forms of resistance that emerge from it. While some scholars (Ferguson, 2015; Harris and Scully, 2015) have argued that the rapid expansion of social grants in South Africa provides the potential for a new politics of redistribution and decommodification, we argue that this has been undermined by technologies which distance recipients from the state and cast them as potential borrowers rather than citizens.

Methods

From 2016 to 2017, we conducted research on the role of financial technology company Net 1 in South Africa’s social grant distribution system. Conducted in collaboration with the International Labour Research and Information Group (ILRIG) in Cape Town, we undertook focus groups with grantees, monitored grant pay points and conducted short interviews with those accessing loans through Net1 subsidiaries at multiple locations. Over a 3-month period in 2017, we visited pay points, shopping centres, banks and loan offices in Gugulethu and Philippi townships in Cape Town. We arrived early in the morning and stayed through the day, observing where people went to collect their grants and interactions they had with those involved in distribution. In 2016, we conducted focus groups in Gugulethu and Philippi townships with grantees, almost all women, who were unemployed or self-employed, often working as informal waste-pickers or home-based carers. Throughout this period, we also attended a number of community meetings, in both Cape Town and rural parts of the Western Cape, where participants spoke about the problems of illegal deductions from the grants. In addition, ILRIG organized a series of workshops with grant recipients from across Cape Town where we delivered presentations on the role of Net 1 and provided space for grantees to provide feedback on their experiences receiving grants and accessing loans, insurance and other financial services. We took detailed notes during all workshops, focus groups and observations. In 2021, during the COVID-19 pandemic, we resumed research on changes to the social grant system, focusing on the role of the national post office in the distribution of pandemic-specific income support programmes. Public health restrictions during this period made any in-person research challenging, so interviews with government officials, non-governmental organization (NGO) staff and community activists were conducted via Zoom and WhatsApp. These were accompanied by a review of government policy documents and online news stories.

Cash transfers and the logics of financial inclusion

The financial inclusion agenda has emerged over the last decade as a powerful force in global development policy (Bernards, 2019; Mader, 2018; Soederberg, 2014; Torkelson, 2020). Financial inclusion is considered a major strategy to achieve the UN Sustainable Development Goals (SDGs) and is increasingly linked to the growth of new forms of social protection policy in the form of cash transfers. In the wake of the 2008–2009 global financial crisis, it has been heavily promoted by the G20 through their Global Partnership for Financial Inclusion (GPFI, 2010), aimed at stabilizing the global economy and providing access to financial services for lower-income markets. As Torkelson (2020) puts it, cash transfers provide an ideal partner to advance these techno-financial fixes. According to one review of 212 cash transfer programmes, 20% are now bundled with at least one financial service or product (Clemence and MacLellan, 2017). This integration is described in development policy circles as government-to-persons (G2P) payment systems. Funded by the World Bank and Gates Foundation, the G2P model is aimed at using digital technologies to accelerate financial inclusion (World Bank, 2020a). Increasingly, this model rests on the use of mobile and digital payment technologies to connect the poor with financial products and services and, ultimately, shape patterns of spending behaviour.

The ascent of financial inclusion to development doctrine builds on the market-based solutions to poverty that emerged in the 1990s, specifically the emphasis on unlocking the various ‘capitals’ the poor possess (De Soto, 2000). The so-called microfinance revolution promised to tap the entrepreneurial potential of the poor and, through short-term credit, allow them to pull themselves out of poverty (Roy, 2010). However, the failure of this movement to make any demonstrable impact on poverty levels, a questionable record on women’s empowerment and spiralling debt levels has led to a wholesale rebranding, if not a critical reflection, on the model itself (Duvendack et al., 2011; Fraser, 2009; Guérin et al., 2015; Mader, 2018; Stewart and et al, 2012). As Mader (2018) argues, financial inclusion seeks to go beyond the limits of microfinance, and it has been supported in this by a vast and well-funded network of development agencies, philanthropies, expert research networks and financial technology companies. Gabor and Brooks (2017) have described this policy network as an FPD complex, which has made significant investments in developing digital payment, registration and banking systems aimed at addressing some of the failures of the micro-credit model.

A recent strategy to overcome some of these failures has been linking cash transfer programmes, now extensive across most developing countries, with digital payment technologies. In 2010, the G20 developed a Financial Inclusion Action Plan aimed at improving levels of financial inclusion among members and non-members. The plan was revised in 2014 and 2017, and in its most recent iteration, includes a specific focus on targeting underserved and vulnerable groups, specifically through the digitization of financial products, services and identity systems (GPFI, 2017). One of the major ways this is to be accomplished is through the digitization of transfer payments as a portal to access other financial services and products. While initially aimed at digitizing cash transfer payments, the G2P model has attracted funding from the Gates Foundation and expanded to include payment infrastructures that provide financial services to recipients (CGAP, 2011, 2019). In 2020, the World Bank and Gates Foundation launched the G2Px programme, aimed at developing digital payment ecosystems that accelerate financial inclusion and fintech innovation in low-income markets. Because cash transfer values are typically low, the goal of financial inclusion policies is frequently to help recipients effectively manage these small sums through financial literacy tools, accessing to savings accounts and credit (Mader, 2018).

The financialization of social policy has been documented by Lavinas (2017) in her work on Brazil under Worker’s Party (PT) rule. Soederberg (2014) makes a similar argument by drawing on evidence from the United States and Mexico, where she details the expansion of credit markets amid crises of social reproduction generated by the dissolution of welfare regimes, stagnant wages and the privatization of public services. In the South African context, Ansari (2022) situates social grants within the broader context of the country’s neoliberal and financialized economy. He argues that social welfare programmes in post-apartheid South Africa have been shaped by ‘the direction of the Treasury in accordance with the interests of international investors’ (Ansari, 2022: 553). As a result, social policy has been constrained by a broadly neoliberal macro-economic environment and the financialization of the state, which has facilitated the integration of the poor into circuits of international finance.

There is growing attention to how these welfare infrastructures are experienced and contested by recipients across the Global South. The expansion of the biometric Aadhaar card has been a critical component in the expansion of social welfare provision in India, and yet for the poor, it has often entailed further waiting and frustration as they are ‘pushed around’ by state bureaucrats. Raphael’s (2022) work on Delhi street traders demonstrates how the process of ‘waiting for welfare’ is shaped by hierarchies of class, caste and gender. Marginalized groups navigate the complexities of waiting using a variety of inventive practices that allow them to gain access to critical livelihood supports. Carswell and De Neve’s (2022) ethnographic work in Tamil Nadu reveals that digital, biometric and mobile technologies are transforming the experience of receiving social protection. Mobile technologies produce new opacities as paper-based forms give way to text messages that can easily be lost. Technologies also produce new forms of exclusion as registrations and payments are subject to errors that can take months to resolve. Registration and payment technologies are ultimately mediated by human agency, as recipients seek assistance from kin and non-kin in order to resolve registration and payment problems. ‘Such mediations’, they write, ‘constitute the vital human infrastructure that makes systems – old and new – work for the poor and that helps to materialize technology’s promised benefits’ (Carswell and De Neve, 2022: 138).

A key focus of scholarship on the technologies of cash transfer infrastructures has been how they are mediated by human and non-human actors. Donovan’s (2015) work on cash transfer payments in Kenya demonstrates how the move towards biometric and digital payment models was rationalized as a way to prevent fraud and to prevent leakage and human error. Yet the adoption of these technologies was hardly smooth, with biometric scanners affected by dust and heat and local shopkeepers and aid workers playing a critical role in navigating and repairing distributive infrastructures. The critical role of local shopkeepers in the distribution of funds highlights the ‘curious mix of private, public and civil society actors who now constitute humanitarianism’ (Donovan, 2015: 20). Shopkeepers were not simply distributors of funds but served as critical intermediaries, fixing technologies and answering technical questions. Contemporary governmentality, Donovan argues, is no longer exercised through government agencies or even aid offices but through socio-technical negotiations of infrastructure on the ground.

While the mediation of these infrastructures can make these systems work for the poor, they are also critical to finance capital’s accumulation strategies. In Kenya, the popular M-Pesa mobile payment and lending system relies on what Donovan and Park (2022) call algorithmic intimacy, in which the family and kinship relations are frontiers of accumulation for M-Pesa’s owner Safaricom. In the context of poverty and stagnating incomes, small loans through M-Pesa allow people to maintain kinship bonds, often across urban and rural space; however, they also create new anxieties and constraints. Similarly, Kusimba et al.’s (2016) work on women and digital payments in Kenya suggests that mobile money transfers are a way to shore up matrilineal ties in the context of significant livelihood changes and the pursuit of economic security through a variety of means. For fintech firms, the family has become both a site of accumulation, through various methods of repayment and data gathering to generate new commercial insights. While these digital infrastructures are mediated by social relations, it has not always been straightforward to convince people to adopt systems like M-Pesa. As Iazzolino and Wasike (2015) argue, cash continues to play a critical role in articulating and mediating social relationships in Kenya, and because it is universally accepted, many are not willing to adopt entirely cashless forms of payment. In the South African context, the dominance of cash among the poor has led financial inclusion advocates to push the government to fund digital payment systems which facilitate the transition to cashless payments and enable credit and risk scoring (FinMark Trust, 2019b).

‘The money is never enough’: financial inclusion amid a crisis of social reproduction

The expansion of social grant spending has coincided with a broad crisis of social reproduction affecting millions of South Africans. Between 2010 and 2015, state spending on social grants increased by 39% with over 45% of households in the country receiving at least one grant (Statistics South Africa, 2016). Over the same period, unemployment hovered around 40% and 28% of the population lived in exreme poverty, defined as below the food poverty line (Statistics South Africa, 2017). For many years, grant increases have not kept pace with inflation, and while individual years saw more sizable grant increases, the real value of the Child Support Grant has declined in the face of rising food costs (Webb and Vally, 2020). Many of the reports on the illegal deductions crisis which framed the problem as a case of predatory lenders alone failed to address the reasons why grantees seek out short-term loans. As Bateman (2014) has argued, rather than a solution to poverty, micro-lending in South Africa has intensified poverty, leading to high rates of over-indebtedness among the poor.

Phiwe is a university student living with her mother and sister in Khayelitsha township. When we spoke, she had just withdrawn R2000 in student loan funding from her bank account. ‘This money is meant for me, but I divide it in half, I use R1000 for toiletries and clothes, and the rest is for the house’. Phiwe’s mother works as a domestic worker, earning just under R4000 a month, around one-quarter of this is set aside for transportation to work. Previously, the household had received 2 child social grants, 1 for each sister, but when Phiwe turned 18, this became 1 grant. Phiwe sets aside some of her own student loan funding to cover household expenses and to repay loans. However, this is never enough. ‘Sometimes we need loans for school uniforms or fees for my sister, those are the big ones . . . That money is never enough, everything costs, and the taxis they eat our money’. Like many South African households, wages alone are insufficient in meeting basic expenses and are supplemented with income from a variety of other sources including student loans, social grants and, when necessary, short-term loans.

Phiwe’s story was a recurrent one throughout our research. While most households did not have access to student loan funding, other forms of borrowing were often seen as necessary to tide households over through periods of financial instability or to afford expenses such as medication, school uniforms and transportation. In all the cases we encountered, loans were not used to further entrepreneurial ambitions, as they are often imagined to, but rather to sustain everyday household consumption. Research has shown that borrowing among grantees is primarily used to finance food (42%), clothing (25%) and transport (12%) expenses (Finmark Trust, 2019a). In the absence of any significant increase in employment, borrowing is driven by a deep crisis of household social reproduction in which wages are either absent or inadequate. As unemployment rates soared in South Africa, household remittances from wage labour have declined in significance and have increasingly been supplanted with redistribution derived from public sources, grants, pensions and state-provided student loans (Ferguson, 2010, 2015). Rather than households being constructed around wage earners, it is increasingly common to find them cohering around those who have regular access grants. Because the vast majority of grants paid each month are Child Support Grants, it is frequently women who take on the role of budgeting and distributing these funds across the household and, ultimately, borrowing and repaying loans. Rather than empowering women, gendered expectations around care-giving and financial support have been associated with household conflicts reflecting unequal power relations within households (Button and Ncapai, 2019).

We observed these dynamics at the Gugulethu Square shopping centre one grant payment day. When we arrived at 5:30 am, recipients, most of them women and some with sleeping children, began placing cardboard and milk crates on the ground outside the shopping centre’s main doors. By the time the centre opened at 7:30 am, the line had swelled to several hundred and traders had begun setting up shop in the parking lot – toiletries and secondhand clothes were displayed, and fires were lit to grill meat. Inside the centre, grantees could choose to withdraw their funds from a bank or from one of the two grocery retailers; the vast majority chose the latter and withdrew most of their funds in cash at the grocery checkout. Over 85% of grant recipients withdraw all their funds in cash during the first 5 days of the month (FinMark Trust, 2019a: 30). After withdrawing their funds in cash, recipients purchased groceries or wares from sellers outside the shopping centre. A pensioner we interviewed lamented that while the R1000 per month he received was enough for food, it never covered water and electricity or demands from other family members. ‘That money, it doesn’t last, by the end of the month we are hungry . . . and we are living 10 in our house’. To address this, he went down the road to the neighbouring Philippi township where a range of loan providers had set up shop.

The Moneyline loan office in the Philipi plaza shopping centre was only visible due to the long line that had formed outside of it. There was no signage on the storefront, just an armed guard and a small sign from the National Credit Regulator stating that the establishment could legally offer credit services. People from the community knew simply by word of mouth that they could access short-term loans here. Borrowers, almost all women or pensioners, accessed loans of between R410 and R1050 to be repaid over a 3- to 6-month period (see Figure 1). Moneyline claimed it did not charge interest, just ‘service fees’, although these service fees essentially acted as a form of interest, in the realm of 30–40% on the principal amount. Borrowers were also subject to a range of service and transaction fees, which varied from R3-R6 for each balance inquiry and cash withdrawal.

OPEN IN VIEWER

Figure 1.

Grantee holds loan document from Moneyline Office, NET 1 subsidiary, Philippi, Western Cape, June 2016.

Potential borrowers entered the office and were asked to present their SASSA card and scan their thumbs using a biometric scanner, Moneyline agents could then pull up their grant and transaction history, perform a credit-worthiness check and sign them up for an EPE card. Those rejected by Moneyline (often due to existing loan balances) were simply directed to a ‘Cash Connection’ loan office around the corner, which prominently displayed a ‘Pensioners Welcome’ sign above the entrance (see Figure 2). Here credit checks were performed using a Net1 UEPS card reader, which allowed the lender to confirm grant payment schedules. Once again, borrowers were not charged interest, but ‘initiation fees’ and monthly ‘service fees’ which, in many cases, resulted in a 50–60% rate of interest on the principal. Access to Net 1’s card readers meant that other credit providers could easily check transaction histories and extend loans. Even those rejected by Moneyline were encouraged to sign up for the EPE card, which would allow them to make payments and purchase airtime and electricity vouchers. For Net 1, this was also a strategy to retain a sizable client base in order to continue to offer grantees loans after the Post Office took over grant payments and ring-fenced accounts preventing automatic deductions. By 2019, just under 1 million grant recipients were still using the Easy Pay accounts for transactions even as CPS was no longer involved in grant distribution (James et al., 2020).

OPEN IN VIEWER

Figure 2.

Grant recipients wait in line to access microloans, Philippi, Western Cape, June 2016.

For many grantees, loans were simply seen as necessary in a context in which grant rates were insufficient and wages were low or absent. In other words, people simply lived with debt as an ongoing feature of their lives, often taking out one loan to pay off another or accessing informal loans to pay off formal loans. Participants reflected that there were some benefits to accessing loans through Moneyline. The money was paid out right away, and the rates of repayment, while high, were often preferable to borrowing from mashonisas in the township. While companies like Moneyline provided easier access to loans, engaging them provoked significant unease among grantees.

Participant 1: Once the money is finished [referring to the Child Social Grant], we have no option. Either we go to umashonisa in our neighbourhood, or we go to Moneyline . . . but they don’t give you a contract, so you get deductions for unknown reasons . . . You’re depending on the child grant, and it’s heartbreaking, it causes a lot of anxiety if you’re not going to get paid on the first of the month.

This statement is revealing in a number of ways. First, it confirms that grantees access loans because they have limited recourse to other income sources, they had no option but to go and seek out loans. Second, borrowing money exposed them to a number of risks related to the terms of the loan contract and, potentially, deductions that would occur without their knowledge or consent. Despite these risks, accessing loans through formal lenders was preferable to informal lenders. Loans were seen as simultaneously necessary to support households and source of profound anxiety. Debt was simply lived with as a condition of low grant values, widespread unemployment and rising costs. As Donovan and Park (2022) have noted in Kenya, digital debts, which are seen as necessary to mediate kinship relations, are ultimately a product of stagnating economy where access to everyday forms of liquidity is needed to make ends meet. Digital debts, they argue, allow for immediacy and liquidity but can also transform social ties and produce new anxieties and constraints.

‘They play us like a soccer ball’: confusion and mediation in social grant infrastructure

In a focus group with grant recipients in Gugulethu, a participant held up an EPE card, most commonly known as the green card. The green card, she told us, was given to her when she went to a Moneyline office to get a loan of R600. Loan officers told her the card would now be her de facto SASSA card, and all her social grant money would be deposited into this account. If she wanted to check her balance, find out whether her grant had been deposited or apply for a future loan, she needed to use the green card. However, when her loan had been repaid in full and the deductions continued, she could not get a clear answer on why the deductions continued through her green card or how to prevent them.

Participant 2: So, I go to Moneyline and they say they will check. They didn’t do that, so I call SASSA and they say no I must come and see them in town. I go there with all of my details . . . And I show them these withdrawals, and my loan statements, I say I want my money. They say no this is a Moneyline problem. So you see, they play us like a soccer ball. They pass you to SASSA they pass you to the bank, to Moneyline, back and forth.

The EPE card was developed by Net 1 in response to the state’s attempt at limiting deductions from grantees’ accounts through their SASSA cards. Both the EPE card and SASSA cards were linked to bank accounts held by Grindrod Bank, which allowed Moneyline to transfer grant deposits between these two accounts through automated transfers. By applying for a loan, grantees consented to have their grants transferred to an EPE-linked account from which interest, fees and loan repayments could legally be deducted. Because grantees were required to sign up for an EPE card in order to access Moneyline loans, Net 1 was able to capture a significant client base which they retain to this day.

This story was a recurrent one throughout our research and was reflected in a range of media reports about the challenges grantees faced in contesting deductions. In one case, a group of pensioners reported recurrent deductions of R300 a month for a Moneyline loan they did not authorize. They could not contest this with local SASSA officials but were redirected to a phone number which was never answered (Damba-Hendrik, 2016). In Torkelson’s (2021) account of illegal deductions in the rural Western Cape, a grantee named Sophia borrowed money from a Net 1 agent who visited her community, but due to similarities to grant payment infrastructure (specifically the suitcase containing the UEPS scanner and card reader), she believed the agent worked for SASSA. When she tried to dispute ongoing loan deductions, she was unsure whether to approach the individual lender, the state or Net 1. Travelling to a Net 1 office in a rural town was expensive and Net 1 agents placed blame for the deductions on other micro-lenders in her community.

Among advocates, mobile and digital cash transfer infrastructures are touted as more efficient, more secure and more empowering (CGAP, 2019; World Bank, 2020a,b). This techno-optimism frequently ignores the complexities of how people engage with payment infrastructures and how these systems can create confusion, anxieties and exclusions. One common issue we encountered in our research was confusion over who was ultimately responsible when something went wrong with grants, specifically when amounts were not deposited on time or when deductions continued well beyond the end of a loan repayment period. In urban areas such as Gugulethu and Philippi, the vast majority of grant recipients withdraw their funds in cash from pay points such as bank ATMs or grocery store checkouts. At a shopping centre in Gugulethu, we spoke with a grantee who had just withdrawn her grants from a grocery store.

Participant 3: I get two child grants so R700, but you see here there is money missing [pointing to a transaction fee of R22.95 for a withdrawal]. They also say I have paid for airtime, but I never paid for airtime. I don’t know who took this money. [pointing to R100 for prepaid airtime].

When she asked how she could find out about these transactions, she was told to call SASSA or to go and get a full transaction history from a bank – which would incur an additional fee. The majority of recipients we spoke with were not provided with any transaction statements, meaning it was often impossible to see who had withdrawn funds, or dispute individual withdrawals. They were directed to call SASSA or to call Moneyline if they were using the Easy Pay Everywhere cards. Rather than efficient and secure, digital payment systems created new opacities, confusion and anxieties as bank accounts were seen as permeable and relationships with the state mediated by private actors.

This mediation of cash transfer payments by a mix of public, private and humanitarian actors has been noted in both the Kenya and Indian cases (Carswell and De Neve, 2022; Donovan, 2015). In the South African case, we see a similar mix of public, private and civil society actors involved. Rather than dealing directly with agents of the state, the majority of grant recipients engage on the first of the month with ATMs, bank clerks and grocery store checkouts. In these instances, these is little opportunity to dispute transactions or lodge complaints. The state was simply not present, obscured behind grocery checkouts, screens or text messages. In one case we encountered, disputing deductions fell to the children of grant recipients:

Participant 4: My mother got a loan with Moneyline, but these deductions they continued, and we went to go and speak with the councillor. He said no you must go and talk to the loan office, so we went there, and it was just stress. We are made buffoons, we can see that there is a the name Grindrod on the back of the SASSA card and the Green Card. When you talk to SASSA, they will say, no we don’t allow loans from SASSA accounts, so you must talk to the green card, so you can see what’s happening. They are fooling us.

Grindrod bank was a critical actor in the grant distribution infrastructure during this period, providing bank accounts for some 17 million grant recipients, and allowing companies like Moneyline to set up millions of additional accounts for those with EPE cards. As this recipient noted, this relationship was clear on the back of cards, but disputing any transactions was laborious and frustrating. They were, as she puts it, ‘made buffoons’ by both government and private companies who collaborated to drain their limited incomes while providing no channels to challenge this. The involvement of recipients’ children in navigating digital payment infrastructures raises an additional concern of digital exclusion. Researchers have warned that digitizing social welfare systems may marginalize clients with limited digital access and skills (Alston, 2019; Melchiori et al., 2023). Within South Africa, government reports reveal that mobile and online registration models, such as those used in the rollout of grants during the COVID-19 pandemic, often exclude those in rural areas with lower levels of education (DSD, 2021). In our experience, the process for applying for and receiving a grant is largely carried out in English, and in the absence of translated documents or support services, it is unsurprising that a variety of kin and non-kin intermediaries have emerged to navigate these infrastructures.

Conclusion

The rapid expansion of temporary social protection measures during the COVID-19 pandemic provided a significant boost to the Fintech-Philanthropy Development complex. For example, the World Bank (2020a, b) saw the pandemic as ‘an opportunity to fast-track changes already in the works such as interoperability, mobile money adoption, and digital financial services’. Initiatives funded by the bank through the pandemic sought to go beyond access to financial services and use machine learning tools and mobile data systems to develop targeted social assistance programmes (Aiken et al., 2020). For proponents, these technologies are neutral systems aimed at improving access to income supports and helping the poor manage their incomes. Yet as the South African case reveals, who owns these technologies, how they are used and what forms of oversight exist over their use matter a great deal. This is not to suggest that previous distributive infrastructures are superior – or that there is no room for technology in the provision of state services – but that the distributive infrastructures that underpin state welfare models are increasingly central to the accumulation strategies of finance capital.

In the South African context, we highlight how these technologies are changing the experience of receiving social protection, limiting the potential these policies could play in advancing transformative and redistributive politics. First, the fusion of financial technologies and cash transfers occurs amid a crisis of household social reproduction characterized by economic stagnation, high unemployment and insufficient social protection. In this context, borrowing is used to maintain kinship relations, ensure basic household income stability and cope with shocks. Financial inclusion is not experienced as a pathway out of poverty, or as a tool to spur entrepreneurial aspirations, but as an increasing necessity in the context of livelihood crises.

Second, the informational and technological infrastructures which increasingly underpin cash transfers in South Africa and across much of the Global South have introduced new forms of mediation, confusion and exclusions. Grantees rarely deal with the state directly, navigating access to critical livelihood supports through a complex web of text messages, online platforms and call centres. In South Africa, as elsewhere, the adoption of these technologies has produced new forms of waiting, mediation and exclusions (Carswell and De Neve, 2022; Chaudhuri, 2021; Donovan, 2015; Raphael, 2022). As the use of these technologies expands, it will be increasingly important to understand their qualitative dimensions, specifically how they affect the nature of social citizenship and how citizens engage with the state.

Third, the outsourcing of grant distribution to private companies, alongside limited investment in public utilities, has weakened the state’s capacity to provide social protection. This has been abundantly clear in the national Post Office’s inability to pay the COVID-19 SRD to the point that grantees have since been directed to banks and grocery retailers to withdraw funds. Rather than an oversight, much of this has been by design. As Ansari (2022) has noted, a more universalist social policy has been restrained by the National Treasury in the interest of maintaining international investment and favourable credit ratings. Lack of state capacity provides significant opportunities for financial technology companies to both benefit from state contracts and, ultimately, shape the direction of social policy.